Friday, August 21st, 2009
We cannot statistically prove anything because the statistics themselves lack the power to do so. Yet people still continue to believe they can – mostly because they don’t know what significance means. What can we do to prevent this lack of knowledge from spreading even further?
Friday, August 7th, 2009
Statistical significance has an outstanding reputation. But does it deserve it? How significant is your knowledge of significance? Take this test to find out!
Friday, May 1st, 2009
Hans Rosling wanted to show the best statistics ever seen at a TED conference. In my opinion, it was only one of the best presentations of statistics. Here are the pros and cons for using animation in data analysis.
Friday, February 1st, 2008
When managers and statistics collide, less is often more. In general, however, there are not enough statistics in data visualizations. The only requirement is that they need to be robust.
Friday, July 27th, 2007
It’s understandable that managers don’t want to spend more time studying a computer screen than they would perusing the business section of the Wall Street Journal to simply get their latest revenue statistics. Yet it is no reason to substitute suitable for foolproof reporting.
Friday, July 13th, 2007
In the days of modern, high-resolution photography, it’s sometimes hard to believe that Galileo’s tools were limited to pen and paper. His observations of the moon are sketched with the precision of a scientist and the artistic prowess of a Michelangelo. His drawings are the data material for his analysis. It’s the artistic prowess of the reproduction – and not the mere ornamentation – that distinguishes first-class information graphics.
Friday, June 29th, 2007
What constitutes statistical thinking? As we take another look at our margarine marketing example, we find that the key is just plain commonsense.
Friday, June 22nd, 2007
Means are often used as the sole representation to characterize distributions for a collective of data. Yet, these types of shortsighted visualization are highly dangerous. How can we present data so that it reveals more than it conceals – and still makes sense to managers?
Friday, June 8th, 2007
If you claim that something is “statistically significant”, your audience will usually be all ears. For many people the term signalizes a type of soothing assurance. We contend that significance is not a matter of coincidence but simply what we wanted to measure in the first place. Misinterpretations of significant results are common. At the end of the day, you need to make your own conclusions. There are no statistical methods that can do this for you.
Monday, May 21st, 2007
German soccer fans don’t often agree on much. Few, however, can deny that the 2006/2007 Bundesliga season was more exciting and suspenseful than it had been for years. Sparklines illustrate the ups, downs and overall drama of the season’s 306 matches.